H. W. To. Optimising the Parallel Behaviour of Combinations of Program Components. PhD thesis, Department of Computing, Imperial College of Science, Technology and Medicine, London, 1995. 24  Home/Search   Document Details and Download   Summary   Related Articles   Check

....the skeletons fold and scan there were static and growing versions. For the static version, the emerging result remains constant in every step. However, for the growing version the size of the emerging result grows by an amount equal to the size of each list element. 2.3.3 To s Approach H. W. To [160] has looked into optimising combinations of algorithmic skeletons. A language for combining skeletons was proposed and a set of primitive skeletons was chosen which is based on the operators of parallel abstract data types 26 (PADTs) 8 . This choice was based on the observation that many ....

H. W. To. Optimising the Parallel Behaviour of Combinations of Program Components. PhD thesis, Department of Computing, Imperial College of Science, Technology and Medicine, London, 1995.

....being performed. More 44 Chapter 1. Introduction complex programs tend to exhibit more complex algorithmic requirements, for example by being multi phased (having more than one significant parallel computation) or by having several levels of useful parallelism. Skeleton work at Imperial College [23, 80] has progressed toward the goal of providing skeletons that are intelligent libraries . These skeletons are composable using special glue skeletons to provide more complex programs. The use of glue skeletons rather than general composition operators allows more careful control of the combination ....

H. W. To. Optimising the Parallel Behaviour of Combinations of Program Components. PhD thesis, Department of Computing, Imperial College, University of London, Sept. 1995.

....on the abstract machine that has been fixed for the construct and on the target architecture at hand. In Rangaswami s HOPP (Higher Order Parallel Programming) model [7] for skeleton oriented programming, nesting of skeletons was limited to three levels and code had to be generated manually. To [8] investigated optimising the combinations of algorithmic skeletons. A language for combining skeletons was proposed and a set of primitive skeletons was chosen. There have also been a number of approaches based on restrictions to the forms parallelism may take. In Blelloch s NESL [4] parallel ....

H. W. To. Optimising the Parallel Behaviour of Combinations of Program Components. PhD thesis, Department of Computing, Imperial College of Science, Technology and Medicine, London, 1995. 8

....by the skeleton nesting in a well organized process graph. Then the process graph needs to be mapped efficiently onto the target machine, balancing the load of nodes and minimizing communication delays. In the literature, different proposals have been made on how these problems can be addressed [12, 19, 23, 17, 2]. In the paper, we discuss an approach which is based on implementation templates. An implementation template [8, 15] is a parametric network of processes recording a good strategy to implement a given skeleton onto a class of target architectures. Each template comes with an analytic performance ....

....unless a large number of entries in the skeleton set is provided. However, libraries with a large number of highly specialized structures are difficult to use and to port on different platforms. Skeleton composition has been studied in the literature by several authors (see for instance [12, 17, 15, 23]) The basic idea of all the approaches is to fix a small set of initial structures and to allow more complex programs to be composed from them. This approach is generally known as structured parallel programming by analogy with structured sequential programming, in which (sequential) programs are ....

[Article contains additional citation context not shown here]

H.W. To. Optimising the Parallel Behaviour of Combinations of Program Components. PhD thesis, Imperial College, London, 1995.

....as a generic higher order complexity function. Also, the models were extended to cover scalability analysis using the isoefficiency function [6] 4 Cost Models for Skeleton Oriented Programming The following researchers: S. Pelagatti [15] R. Rangaswami[17] D. Skillicorn [18] and H. W. To [21] have developed cost modelling systems in conjunction with their models for skeleton oriented programming. Before outlining the cost models they have developed for their skeletons, we need first to present their methodologies for skeleton oriented programming. The skeletons in this section differ ....

....side and the right hand side. For this identity both sides do have equal costs (cost neutral) Therefore, we can label it in the following way: f ffi g) f ffig Other identities were labeled in the same way and used to guide transformations for cost effective implementations. H. W. To [21] has looked into optimising the combinations of algorithmic skeletons. A language for combining skeletons was proposed and a set of primitive skeletons was chosen. The primitive skeletons were based on the operators of parallel abstract data types (PADTs) 7 . This choice was based on the ....

H. W. To. Optimising the Parallel Behaviour of Combinations of Program Components. PhD thesis, Department of Computing, Imperial College of Science, Technology and Medicine, London, 1995.

....2 D torus, linear array and tree. The cost model determines the cost of executing a recognised function on a given architecture. In her system nesting of skeletons was limited only to the first 3 levels and it was not automated as the code had to be generated manually. The work of H. W. To [18] was about optimising the combinations of algorithmic skeletons. A language for combining skeletons was proposed and a set of primitive skeletons was chosen. The primitive skeletons were based on the operators of parallel abstract data types (PADTs) 2 . This choice was based on the observation ....

H. W. To. Optimising the Parallel Behaviour of Combinations of Program Components. PhD thesis, Department of Computing, Imperial College of Science, Technology and Medicine, London, 1995.

....and the code had to be generated manually. 2 All of the recognised functions work on lists. Merge Sort Speedup on Fujitsu AP1000 0 2 4 6 8 10 12 14 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 No. of Processors Fig. 9. Merge Sort Speedup Results on Fujitsu AP1000. The work of H. W. To [23] was about optimising the combinations of algorithmic skeletons. A language for combining skeletons was proposed and a set of primitive skeletons was chosen. The primitive skeletons were based on the operators of parallel abstract data types (PADTs) 3 . This choice was based on the observation ....

H. W. To. Optimising the Parallel Behaviour of Combinations of Program Components. PhD thesis, Department of Computing, Imperial College of Science, Technology and Medicine, London, 1995.

....effective optimisation can be achieved, but it requires detailed analysis of the target architecture and makes compilation time consuming. The technique presented here is expected to replace the present optimisation algorithm. A number of global optimisation approaches have also been tried. To [10] gives a polynomial algorithm to choose among block and cyclic distributions of data for a sequence of data parallel collective operations. The implementation of each operation is fixed by the choice of data distribution. The complexity of the algorithm is n 2 a where n is the number of steps ....

H.W. To. Optimising the Parallel Behaviour of Combinations of Program Components. PhD thesis, Imperial College, 1995.

No context found.

H. W. To. Optimising the Parallel Behaviour of Combinations of Program Components. PhD thesis, Department of Computing, Imperial College of Science, Technology and Medicine, London, 1995. 24

No context found.

H. W. To. Optimising the Parallel Behaviour of Combinations of Program Components. PhD thesis, Department of Computing, Imperial College, London, U.K., Sept. 1995.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC